ipv6: Handle route deletion notification
For the purpose of route offload, when a single route is deleted, it is
only of interest if it is the first route in the node or if it is
sibling to such a route.
In the first case, distinguish between several possibilities:
1. Route is the last route in the node. Emit a delete notification
2. Route is followed by a non-multipath route. Emit a replace
notification for the non-multipath route.
3. Route is followed by a multipath route. Emit a replace notification
for the multipath route.
In the second case, only emit a delete notification to ensure the route
is no longer used as a valid nexthop.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index f1535f1..b579fae 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -487,6 +487,7 @@ int call_fib6_multipath_entry_notifiers(struct net *net,
struct fib6_info *rt,
unsigned int nsiblings,
struct netlink_ext_ack *extack);
+int call_fib6_entry_notifiers_replace(struct net *net, struct fib6_info *rt);
void fib6_rt_update(struct net *net, struct fib6_info *rt,
struct nl_info *info);
void inet6_rt_notify(int event, struct fib6_info *rt, struct nl_info *info,